Connections +
Feature

Extreme Ethernet

Its use in industrial settings is growing, largely because organizations are walking away from traditional proprietary topologies.


November 1, 2003  


Print this page

One of the most common data communication protocols used in enterprise applications is Ethernet. Developed in the 1970s by the Xerox Palo Alto Research center, it was used as LAN technology for offices.

Standardized in 1983 by the IEEE (802.3 standard), Ethernet’s popularity has led to the development of even higher speeds.

Its advantages are becoming applicable in industrial environments, the result of a more open standard for data acquisition and transmission is required for industrial automation.

An increasing number of companies are considering Ethernet rather than the traditional proprietary topologies such as Fieldbus, Modbus, Interbus and Profibus.

The use of Ethernet protocol is increasing in process control, building automation, traffic control systems, power stations, medical, and wastewater treatment applications.

What is giving rise to its use? Ethernet provides numerous advantages to office environments, allowing workers to share files, printers, search the Internet, and support high bandwidth applications.

A factory floor has additional requirements, including complex communication needs such as accessing data from workstations, I/O devices, and other automation systems. Factory floor data is time-sensitive and requires real-time communication.

Historically, industrial automation protocols were proprietary and locked users into a specific architecture. Ethernet’s performance, low cost, and its standardized PC and Windows compatibility make it attractive for industrial applications.

Ethernet capability is being built into industrial measurement equipment such as the I/O devices and data acquisition equipment used to interface directly with thermocouples, strain gauges, flowmeters, etc.

Ethernet challenges

Interoperability or interchangeability between devices from different vendors is an issue for Ethernet in industrial environments. With regards to the OSI model, Ethernet only provides physical and data link protocols. The upper layer protocol determines which devices can connect and inter-operate at the network layer.

Determinism, the ability to predict when information will be delivered, needs to be addressed.

Originally, Ethernet was half-duplex and existed on a bus topology. In shared Ethernet, all the network users share one collision domain.

In half-duplex environments, the network access is controlled by CSMA/CD (Carrier Sense Multiple Access with Collision Detection). Each device on the network senses whether the line is idle.

If the network is idle the device begins to transmit data, other devices can be transmitting at the same time. Collision occurs when two or more devices are transmitting. Consider an industrial scenario in which a robotic arm solders a component, but a congested network causes the robotic arm to fail.

Still, several improvements have been made to Ethernet, making it advantageous to industrial automation:

They include:

Segmentation – subdividing the collision domains;

Higher bandwidths–the development of Fast Ethernet, Gigabit and 10 Gigabit technologies;

Switched Ethernet, and

Interoperability

Switched Ethernet separates collision domains into point to point connections between the network components and equipment, allowing full bandwidth availability for each connection. The separate pair of wires used to detect collision is now used for transmission, increasing transport speeds.

Several proprietary and open fieldbus networks (application layer) are being used in industrial automation applications. The most common control networks are Profibus, DeviceNet, ControlNet, and Foundation Fieldbus. These networks include standardization at the application layer and provide a higher level of interoperability. Ethernet is being considered for these control networks as well.

The advantage of Ethernet is its wide spread acceptance, cost and speed. Industrial protocols competing for acceptance include Ethernet/IP, PROFInet, IDA, and Foundation Fieldbus.

A detailed analysis of Ethernet protocols is beyond the scope of this article. However, Ethernet/IP is considered a front runner due to its popularity and sponsorship by The Industrial Ethernet Association, the Open DeviceNet Vendor Association (ODVA), and ControlNet International (CI).

Ethernet/IP is an industrialized version of Ethernet TCP/IP, which uses TCP/IP encapsulation to provide a common application layer protocol over Ethernet designed to handle both implicit and explicit messages.

The standard application layer allows interoperability and interchangeability among industrial automation and control devices. It does this by using both the DeviceNet and ControlNet standards called Control and Information Protocol (CIP).

It defines the access, behavior and extensions, allowing different devices to be accessed using a common protocol. Using the CIP layer over Ethernet/IP offers consistent devices access. It organizes network devices as a collection of objects and means it can use one configuration tool to configure disparate devices on a network.

Industrial considerations

Several considerations are being debated, the resolution of which will determine the future of Ethernet within the Industrial market. A few of the issues are outlined below.

SNMP (Simple Network Management Protocol) was developed in the 1980s to provide access to remote devices.

Today, it is used to manage a wide range of network devices, including industrial I/O devices, allowing network managers to monitor the health of the network. The importance of SNMP will increase as more industrial devices adopt Ethernet and Internet protocols.

Within the networking industry the functionality of the TCP/IP protocol and Ethernet are closely aligned. Ethernet devices have appeal because of their ability to incorporate TCP/IP and its application services.

Manufacturers, suppliers and integrators are analyzing Ethernet and TCP/IP to support the information, speed, and applications manufacturing will require in the future. There has been a proliferation of Ethernet TCP/IP application layers in various protocols such as Modbus/TCP, Ethernet/IP and ProfiNet, which continues to drive Ethernet’s acceptance in the factory.

Dust, temperature extremes, moisture, and other outside factors affect industrial applications and cause chaos with equipment. If not using an enclosure, IP ratings for dust and water need to be considered. The IP classification system is the degree of protection provided by an enclosure against objects or water.

Elements such as moisture, dust, and debris are eliminated with the use of equipment enclosures. In North American, NEMA, UL, and SCA are common rating systems used to measure enclosure performance. IEC ratings are used in Europe.

The typical connection for Ethernet may corrode; wear, or clog with debris causing failure. Connectors meeting IP67 criteria follow three different approaches: RJ-45, the M12 connector, and a hybrid connector based on RJ-45 technology.

The RJ-45 connector is sealed in an IP67 housing unit, the M12 connector has either a four pin or eight-pin option, and the hybrid is an RJ-45 connector with additional contacts for power distribution.

The M12 connector is already used in industrial applications, specifically the automotive industry. The connector is the standard for sensor connections and some fieldbus systems. The RJ-45 is common in Ethernet office systems, but not acceptable for industrial environments without some type of seal around the connector.

Network administrators need to determine if their environments require conformal coating on network devices. Conformal coating is sprayed onto components (circuit boards) to protect it from moisture, fungus, dust, corrosion, abrasion, and other environmental factors. Enclosures would eliminate the need for conformal coating.

The Role of media conversion

The reality of a network managers’ world is the challenge of creating and maintaining a state of the art data network with a limited budget. Inevitably, network administrators must contend with a variety of protocols, s
peeds, and media in their networks.

Media conversion technology was developed to address these problems. It has evolved from a stop gap technology into a technology that offers network administrators new choices for deploying fiber optics into their networks cost effectively.

Conversion technology enables connections of disparate media types in networks. Media converters are commonly used to connect UTP, copper cabling and fiber optic cabling in a network cabling plant.

They are also used in networks that have legacy cabling such as Coaxial or Twinaxial cabling and need to be integrated with UTP or fiber optic cabling. Converters exist in a variety of form factors, such as standalone, multi-port, and modular chassis to address the various applications that exist.

Media converters are protocol specific; meaning an Ethernet converter is needed to convert 10BaseT to 10BaseFL. Media converters do not convert protocols, such as Serial to Ethernet. However, they exist for a range of protocols including Ethernet, Fast Ethernet, Gigabit Ethernet, T1, DS3, RS232, RS485, V.35, analog phone lines, video, and more.

Conversion products offer the ability to add new devices without replacing costly equipment and cabling, extend network distances by adding fiber where and when needed, provide a link between media types, transparent to the network, keep pace with growing demand and new technology and integrate high bandwidth devices into the network.

Janel Ryan is the market manager with Transition Networks in Minneapolis, Minn. She can be reached at janelr@transition.com.


Print this page

Related